Are you using SCAN?
What is connection string which you are using to connect on database?
Yes, I am using scan.
$ tnsping DB_BALANCE
Used TNSNAMES adapter to resolve the alias
Attempting to contact (DESCRIPTION= (ADDRESS_LIST= (LOAD_BALANCE=YES) (FAILOVER=YES) (ADDRESS=(PROTOCOL=tcp)(HOST=exa-scan)(PORT=1521))) (CONNECT_DATA= (SERVICE_NAME=DB.WORLD)))
OK (10 msec)
PS: I have another issue with the forum. I am using firerox and not able to copy-paste anything here. :-(
As I have mentioned, the configuration was running fine with scan. It is only causing issue, after I shutdown the first two instances.
Thank you for your contineous interest.
The scan is working fine and I have already cross verified the setup.
Most likely the failover is not working at times. As two of the instances are down, very rarely I am hitting the error but that is stopping me to bringing up the application.
Please let me know if you require any further information.
Please don't take this wrong, but if connection is not working properly the setup is not configured properly or can be some bug.
Are two always to try find out what is happening.
You post configuration of your database ... such as remote_listener,local_listener and services registred in all listener, etc
Or you can open a service request with Oracle Support that will help you with more effectiveness.
Did you tried connect (from host where issue is happening) using vip of each host in each instance where is failing? (i.é bypass SCAN)
Nothing to take it wrong.... I am here get help from all. :-)
remote_listener = exascan:1521
local_listener=DB [Running on 1525 port]
For local listener, DB.WORLD and local instance are the service which are handled..
Try bypass SCAN connecting directly on each host using VIP and Local Listener
That will work as I have already tried that, but I need to use scan.
I am constantly hitting this error and appriciate, if you can guide me.
1. Is there any way to trace out more during SQLPLUS login from application node.
2. By any chance, do I need to keep the local listener up where the instances are down.
3. As DB1 and DB2 (i.e., first two instances of the RAC) are down, do I need to execute any extra step apart from shutting down the instances.
PS: I am not able to do copy-paste here. I am using IE9. If possible, please let me know if this is an expected feature.
Please check the tnsnames.ora file and see if you have an entry where the connect descriptor itself is <scan name>:<port>
If so, then create a new entry in tnsnames.ora as follows and update the remote_listener parameter
(ADDRESS = (PROTOCOL = TCP)(HOST = <ip1>)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = <ip2>)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = <ip3>)(PORT = 1521))
Once the above entry is added to tnsnames.ora file on all the nodes of the database, then do the following
alter system set remote_listener='REMOTE_LISTENERS_SCAN' scope=both sid='*';
alter system register;
Once the above is done retry the connections and verify if you are still facing the same issue or not
Vandana - Oracle